Speech feedback for touch-keying

Abstract

A person keying an unfamiliar number listed in a directory, a driver keying, a mobile telephone, and a blind person making a call all face a common problem: how to key a telephone number without looking at the phone. We studied whether touch-keying that is, keying without looking would be facilitated by a talking keyset which spoke the names of the keys as they were pressed. To establish a baseline level of touch-keying performance, we also studied touch-keying with the standard TOUCH-TONE“ feedback. We found that touch-keying was much more accurate with speech than with tone feedback. Speed was not sacrificed to achieve higher accuracy; on the contrary, touch-keying was slightly but consistently faster with speech than with tone feedback. Speech feedback enabled users to catch and correct, via a CLEAR key, over 80% of their errors; tone feedback was much less effective in this regard. No training or practice was necessary to reap the benefits of speech feedback; moreover, the benefits persisted even after extensive practice. The pattern of errors suggests strongly that mistakes in touch-keying were due to misaiming and hitting the wrong key. Misaiming errors were especially common between keys in the 456 and 789 rows, and between the 8 and 0 keys on the telephone keyset.